4.1 Technical Skills
- Strong experience designing enterprise-scale AI reference architectures across experience, agentic core, retrieval,
orchestration, integration, models, data platform, and infrastructure layers.
- Deep understanding of predictive AI, generative AI, foundation models, agentic AI, RAG, vector and hybrid search,
semantic layers, ontology management, knowledge graphs, enterprise search, and governed knowledge
management.
- Experience defining AI guardrails and responsible AI controls including content filtering, prompt-injection
defense, PII detection and redaction, policy enforcement, safety classifiers, red-teaming, evaluation gates, audit
logging, and traceability.
- Strong knowledge of integration and connectivity patterns including API gateways, service mesh, event and
message streaming, connector catalogs, secure identity propagation, schema validation, traffic governance,
quotas, and protocol mediation such as MCP and A2A.
- Understanding of model serving and lifecycle capabilities including model catalogs and registries, commercial and
open-source model hosting, model routing and abstraction, inference infrastructure, fine-tuning, prompt
libraries, benchmarking, and continuous evaluation.
- Experience with MLOps, LLMOps, and AgentOps practices including pipeline orchestration, model and prompt
registries, CI/CD for models and agents, deployment and rollback, monitoring, drift detection, quality monitoring,
and cost and usage metering.
- Strong Azure architecture skills across Azure landing zones, subscriptions, networking, private connectivity,
identity and access management, policy, monitoring, cost management, container platforms, AI services, data
services, and secure cloud deployment patterns.
- Strong on-premises architecture skills across data center hosting, virtualized infrastructure, container platforms,
GPU and accelerator capacity, storage, network segmentation, secrets management, secure connectivity, backup,
resilience, patching, and operational controls.
- Ability to design hybrid AI deployment patterns spanning Azure and on-premises environments, including
workload placement, data residency, latency, secure interconnect, identity federation, private endpoints, key
management, monitoring, and failover considerations.
